unlimited d~staibution i+ national defence defense ... · particular importance in rectifying strip...
TRANSCRIPT
. .
UNLIMITED D~STAIBUTION
I+ National Defence Defense Nationale Research and Bureau de Recherche Development Branch et Developpment
TECHNICAL MEMORANDUM 86/ 206 J anuary 1986
tl .. . , ... -I •
• !.. ..
SOME WARSHIP SLAMMING INVESTIGATIONS
W.C. E. Ne tnercote M. Mac Kay B. Menon
Defence Research Establishment Atlantic
Canada
Centre de Recherches pour Ia Defense Atlantique
: )
I .. . ..
DEFENCE RESEARCH ESTAIUSHMENT ATlANnC 9 GROVE STREET p 0 80)( 1012
DARTMOUTH , N S .
ezv 3~7 TEI..EPHONE
cooz1 • z6.3IOO
I
CENTRE DE RECHERCHES POUR LA DtftNSE ATLANTIQUE 0 GROVE STREET C . P 1012
DARTMOUTH, N E B2Y 3Z7
' •
. .
UNLIMITED DISTRIBUTION
I+ National Defence Defense Nationale Research and Bureau de Recherche Development. Branch et Developpment
SOME WARSHIP SLAMMING INVESTIGATIONS
W.C.E. Nethercote M. MacKay B. Menon*
January 1986
Approved by T. Garrett Oirectcr/Technclcgy Division
DISTRIBUTION APPROVED BY 7~··· 0/TO
TECHNICAL MEMORANDUM 86/206
Canada
Defence Research Establishment Atlantic
i
Centre de Recherches pour Ia Defense Atlantique
*ARCTEC CANADA LTD.
ABSTRACT
Excessive slamming is the most common cause of speed reduction for frigates and destroyers in heavy head seas. DREA's early interest in frigate/destroyer slamming was limited to the development of computer programs for ship seakeeping performance prediction. Published bottom slamming algorithms were adopted and refined ~o provide a slamming prediction capability for DREA software. In follow-on work described herein, both two-dimensional numerical simulations and three-dimensional model tests were performed by contractors to obtain better physical insight and expand the empirical data base. The two-dimensional numerical simulation results required considerable smoothing to reduce numerical noise, but still yielded form factors in satisfactory agreement with two-dimensional theoretical and experimental results. The simulations also gave unique insight into the girthwise development of slamming pressures. The three-dimensional model tests indicated that pivoted drop tests in waves may be substituted for conventional seakeeping tests, but the general applicability of the results was limited by difficulties with measurement of relative impact velocity. Notwithstanding the velocity measurement difficulties, the three-dimensional tests provided important information on longitudinal and girthwise pressure pulse velocities. Further experimental work will be required to develop a usable experimental data base.
ii
.,
.;, •,
Un tapement excessif est la cause la plus commune de la reduction de vitesse des fregates et destroyers par fortes mers debout. Les premieres recherches du CRDA sur le tapement des frigates et destroyers se limitaient a la mise au point de programmes d'ordinateur pour la prevision du rendement des navires en termes de tenue a la mer. Des algorithmes publies de tapement ont ete adoptes et raffines de maniere a permettre la prevision du tapement au moyen du logiciel du CRDA. Lors des travaux complementaires decrits, des simulations numeriques bidimensionnelles et des essais tridimensionnels sur modeles ont ete effectues par des entrepreneurs afin d'eclairer !'aspect physique et d'etendre la base de donnees empiriques. Un lissage considerable des resultats de simulation numerique bidimensionnelle etait necessaire afin de reduire le bruit numerique, mais on a tout de meme obtenu des facteurs de forme concordant d'une maniere satisfaisante avec les resultats bidimensionnels theoriques et experimentaux. Les simulations ont de plus fourni un aper~u inegale de la propagation des pressions de tapement suivant le perimetre. Les essais tridimensionnels sur modeles ont egalement indique que l'on pouvait substituer des essais de chute avec pivotement dans les vagues aux essais classiques de tenue ala mer, mais que l'applicabilite generale des resultats etait limitee par des difficultes de mesure de la vitesse relative d'impact. Malgre ces difficultes de mesure de la vitesse, les essais tridimensionnels ont fourni des renseignements importants sur les vitesses longitudinale et suivant le perimetre de !'impulsion de pression. D'autres travaux experimentaux seront necessaires pour !'obtention d'une base de donnees experimentales utilisable •
iii
1.
2.
3.
4.
5.
6.
7.
8.
TABLE OF CONTENTS
ABSTRACT
TABLE OF CONTENTS
NOTATION
INTRODUCTION
DREA'S BACKGROUND WORK
2.1 Early Slamming Algorithms 2.2 Improved Slamming Algorithms
IDENTIFICATION OF REQUIRED WORK
COMPUTER SIMULATIONS
4.1 Background 4.2 Test Cases 4.3 Numerical Model 4.4 Discretisation Noise 4.5 Surge and Section Dynamics 4.6 Impact Pressure Distribution 4.7 Form Factor
EXPERIMENTAL WORK
5.1 Drop Test Programme 5.2 Seakeeping Test Programme 5.3 Results
5.3.1 Drop Tests 5.3.2 Seakeeping Tests
SIMULATION/MODEL TEST COMPARISONS
CONCLUDING REMARKS
ACKNOWLEDGEMENTS
REFERENCES
iv
ii
iv
v
1
1
2 3
6
7
7 8 8 9
10 12 13
16
16 18 21 21 26
27
30
31
33
.I,'
·:. !
...;/
'J
"'
NOTATION
A, B, c, K
--:C
h
h
k
k'
~
L
p
Ph
PMAX
Pt
P( •• •)
s
t
T
A v
v
vh
vo
von
Vot
Vv
X
x
subscripts for pressure transducer locations
-assumed- speed-of- sound -in water
half-siding
time, hours
form factor
corrected form factor, per equation 12
grid dimension for numerical simulation
length between perpendiculars
impact pressure
most probable extreme pressure ~n h hours
maximum slamming pressure
most probable impact pressure in t seconds
probability of occurrence
girth
time, seconds
draft
threshold velocity
vertical ship/section velocity
horizontal component of wave velocity
initial impact velocity
velocity component normal to wave
velocity component tangential to wave
vertical component of wave velocity
undisturbed waterline corresponding to y
x corrected for surge
v
X, Y, Z hull coordinate system
y penetration
a buttock angle ----- -- --------- ------ -----~--------- ----------
a
B
e
1-1
z.;A
p
Note:
parameter for extreme pressure probability, Figure 4
deadrise angle
wave slope
surge ratio
wave length
wave amplitude
body-fixed coordinates, Figure 8
density of water
RMS relative motion
RMS relative velocity
pitch angle
The symbols h and a both represent two parameters, but in each case the symbol is in conventional use, so the duplication was retained. The appropriate definition is made clear in the context of each application.
vi
I
"
1. IBTRODUCTIOB
For warships of frigate/destroyer size, excessive slamming is the -~most -c~ormnon·reason-·fo~r- spe-ed~reducTi-on inheavy seas;·--fnus-, aprincipar
goal in hull form research and development is the improvement of slamming characteristics. An implicit requirement for such improvement is the · development of accurate slamming prediction capabilities.
DREA's early interest in slamming was linked to the development of computer programs for ship seakeeping performance prediction. These programs, of the common frequency domain, linear strip-theory type, adapted existing methods to provide predictions of bottom slam pressures and total slam force.
Notwithstanding development of these prediction capabilities, it was recognized that there was still a requirement for more fundamental information on bottom slamming, so further work was conducted, both analytical and experimental. In the first instance, the commercially available computer code PISCES 2DELK was used to simulate drop tests of 2-D wedges, truncated wedges and ship sections. In the second approach, ARCTEC Canada Limited were contracted to conduct an experimental program.
The computer simulations provided tinie histories of local pressures, total force and section dynamics, as well as the development of the girthwise pressure time history. After smoothing to reduce numerical noise, the data were used to derive form factors, which satisfactorily agreed with two-dimensional experimental results.
The experimental work was directed towards development of a data base containing slam-related quantities for warship hull forms. Three warship hull forms, with 'U', 'normal' and 'V' bow sections were used in vertical and pivoted drops, in waves as well as in still water. Seakeeping tests were also carried out with the V bow form. Measurements included local peak pressure, pressure rise time, pulse duration and frequency, for up to 24 transducers.
This memorandum describes the above work and summarizes earlier slamming research at DREA. The two-dimensional simulations have provided insight int.o the fundamental details of the process, but it is clear that three-dimensional model data are essential to the achievement of realistic ship slam pressure predictions. Experimental techniques must be refined before a useful data base can be created.
2. DREA 1 S BA<XGROURD WORK
The catalyst for slamming research at DREA was the development of an in-house computer program for predicting ship motions. PHHS1 (Pitch
1
and Heave in Head Seas) employed conventional strip theory for the calculation of vertical plane motions in unidirectional head seas. This coding was subsequently used together with methodology proposed by Salvesen, Tuck and Faltinsen2 and enhanced by Schmitke3 to develop a 6
····· · degree- o:f--f.reedom- p-re d-iet-ion-capabi-lity.. . ..... -··--·
The utility of PHHS was greatly increased by the addition of algorithms for added resistance, relative motion corrections (wave profile and dynamic swell-up), deck wetness, slamming pressures and human tolerance to vertical motion. Some of these capabilities were unique among contemporary computer programs. Relative motion corrections were of particular importance in rectifying strip theory's typically poor prediction of deck wetness; for example, see Schmitke's contribution to the discussion of Reference 4.
2.1 Early Slamming Algorithms
The first slamming calculations were based on the statistical theory of Ochi and Motter5 together with Chuang's6 experimental impact data. In the conventional manner, slamming was taken to be dependent upon keel emergence and the reaching of a threshold relative velocity on impact. The probability of keel emergence is:
P(keel emergence) = P(keel) = exp [ -~ /2cr2 RM1 (1)
where T is draft and crRM is root mean square relative motion, corrected for dynamic swell-up. Slam probability is then
"'2 2 P(slam) = exp [-v /2cr Rv1 P(keel)
where crRv is RMS relative velocity and the slamming threshold velocity, v, is given by
v = 0.0195/gL/(0.03 cot 8 + h/B)/2
( 2)
(3)
where L is the LBP, 8 is local deadrise angle, h half-siding and B beam.
The probability of slamming is of rather academic interest compared to the consequent slamming pressures. As per Ochi and Motter, PHHS calculated the most probable peak impact pressure in t seconds:
(4)
The section form factor, k, was obtained from a statistical fit to Chuang's data as shown in Figure 1:
k = 1 + (1-exp(-5 8)) (rr/2 cot 8)2 ( 5)
The above expression is a reasonably good model of the experimental data for 8 > 5°.
2
...
""ILl 0 u
ILl a:: ::::> en en ILl a:: a.. 1-~ a.. ::e
SPHERE • ---------- ---ELLIP-SO-ID___ ---------
CURVED WEDGE- 6 2 D WEDGE + CONE X
•
Figure 1 Chuang's Impact Pressure. Coefficient
2.2 Improved Slamming Algoritbms
The early slamming algorithms were open to question through their use of prism impact data, as suggested by comparative work by Ochi and Bonilla-Norat7 and Schenzle et al8
• The reply to discussion of Reference 5 dealt with the large magnitude of 2-D section impact pressure
3
coefficients, or form factors, at some length. Order of magnitude differences between drop and seakeeping test derived form factors were often observed, especially at low deadrise angles.
~From -ana-lysis-of mode-1-~and ful-l-scale seakeeping data, Ochi and Motter chose a threshold velocity of 12 ft/sec for a 520 ft ship. Through Froude scaling this generalised ~o
shape.
~ = O. 0927 ( gL)1 /
2 (6)
This exfression may be criticised for its independence of section Schmitke proposed an alternative derived from Ochi's work:
~ = O. 3 77 (gL/k) 1 /
2 (7)
which was obtained by multiplying equation (6) by (16.56/k) 1/
2; 16.56
is the k value for Station 3 of Ochi's Mariner hull-form, used in the derivation of equation (6). In this way, a threshold impact pressure can be defined:
p ~ (p/2)k ~ = 0.0711pgL (8)
Schmitke noted that equations (7) and (3) produce comparable results for V-bowed warships.
In modifying the algorithms, the very apparent conservatism of prism impact data was recognized and the Ochi-Motter predictor based on experiments with Mariner models10 was adopted for cases with known section offsets. A simplified option allowing the specification of deadrise angle only was also developed using the routines of Reference 10 to generate form factors for a series of truncated wedges (Figure 2) where
T = 15 ft
h = 0.5 ft.
For B less than 5 degrees, k was taken as 30. Since at 6 of 21 degrees the Ochi-Motter and Stavovy-Chuang11 methods produced similar results, the results of the two methods were coupled together as shown Ln Figure 3. Further reference will be made to this combination later.
Slamming probability is of little value itself, so the Ochi-Motter method was again used to calculate the most probable extreme pressure in h hours of operation:
( 9)
4
,*''
__ I I ~
rl-------------==/(_~ECTION t:; I // SHAPE
I 1- 2h _J
Figure 2 Truncated Wedge Approximation
30~~~--~----~--~----~--~----~
a: 20 ~ C,)
~ :E a: 0 u.. 10
0
OM METHOD
10 20
+,x COMPUTED FROM OFFSETS
30
Figure 3 Form Factor For Truncated Wedges
5
z u..
>-1-u; z I.LJ 0
a:) 0 a: a..
PRESSURE
Figure 4 Slatmning Pressure Probability Density Function
Figure 4 shows that the probability of exceeding Ph is too high to be an acceptable design value, so an extreme pressure is calculated:
where Ph(a) is the extreme pressure whose probability of being exceeded in h hours is a. Note that both Ph and ph(a) are independent of V. Reference 4 regards 0.01 as a reasonable value for a.
3. mERTIFICATIOB OF REQUIRED WOK
A number of points became apparent in evaluating the literature, the most important being:
that seakeeping and drop tests predict inconsistent slam pressures;
that full-scale and model-scale seakeeping tests yield consistent slam pressures; and,
(lO)
- that the available seakeeping-based algorithms are derived from limited data, the bulk of which is representative of merchant ship rather than warship design practice.
6
41
... ·
In view of these deficiencies ARCTEC Canada Ltd. was contracted to
carry out a comprehensive series of experiments at model scale including three-dimensional drop tests and seakeeping tests. The contract called for:
- a state of the art survey of experimental methods in slamming;
- an evaluation of current capabilities of experimental facilities;
an experimental program to acquire slamming data for warship hulls; and
- development of a preliminary semi-empirical prediction method.
K
EXTREME V
MODEL No.267
K
NORMAL
MODEL No. 264
I I I I
K
EXTREME U
MODEL No. 266
Figure 5 Pressure Transducer Locations on Bow Forms
The experimental program made use of hull forms developed by the National Research Council of Canada's Institute of Marine Dynamics (IMD)
for a fast surface ship methodical series.12 Normal, extreme-V and extreme-U designs shown in Figure 5 were selected for experiments.
In parallel with the experimental work, numerical simulations were performed to examine some of the fundamental features of ship slamming.
4. COMPUTER SIMULATIONS
4.1 Background
In order to gain a greater insight into the physical parameters
affecting the slamming process, a series of computer simulation cases was run. Computer code PISCES 2DELK (a trademark of Physics International Company, California) was selected for the work. It is available in Canada on the CDC CYBERNET system. Individual cases were set up and run by CDC CYBERNET. The resulting data were transferred to DREA for reduction and analysis.
7
PISCES 2 DELK is a two-dimensional finite difference code with Eulerian, Lagrangian and shell-structure capabilities. It was designed to simulate high velocity impacts and explosions, i.e. over time scales shorter by up to severaL orders of magnitude tha11 tgose characterising a ship slam. The implications of this to the present work are noted in Section 4.3. The methods employed by PISCES 2DELK are described in detail in Reference 13.
4.2 Test Cases
Each case run differed in geometrical parameters and in initial impact velocity. Three, two-dimensional geometrical forms were tested: simple wedges, wedges truncated at the keel to represent half-siding, and forward sections (here referred to as frigate sections) from the 'normal' model No. 264. For wedges, the geometrical parameter was deadrise angle, 8, which could take a value of 10, 20, 30, or 40 degrees. Sections 3, 4, and 5 (O at FP and 20 at AP) were used from model 264. Full scale impact velocity was 10, 15, 20 or 30 ft/sec.
4. 3 Numerical Model
Simulations were run using full scale dimensions. The section fell under the influence of gravity, simulating a free drop tes.t. Figure 6 illustrates the initial (time = zero) geometry for a simple wedge. The water, and a void region above it, were discretised in an Eulerian (fixed) grid. The void allowed upward deformation of the free surface. The
VOID -+--+--!---+--REGION
Figure 6 Initial Geometry for Numerical Simulation
8
t[
...
impacting section was assumed to be rigid, its boundary defined by a
Lagrangian (moving) grid. Standard material properties were modified to
ensure section rigidity and to reduce the speed of sound in water, as noted below.
The equations of motion were solved explicitly in time. In order
to ensure stability of the solution, the time step ~t was restricted to a fraction of Q. /c, where R. is the minimum cell dimension in the Eulerian grid and c is the assumed speed of sound in water. To achieve a reasonable value for ~t, givent, the assumed speed of sound was reduced from the true value, with the limitation that the local Mach number nowhere exceed 0.1, so that the effects of compressibility on impact dynamics could be neglected.
4.4 Discretisation Noise
The data were generally quite noisy, as illustrated by Figure 7.
Each curve on the figure shows pressure as a function of girth at a time indicated by the right-hand vertical scale. For low deadrise cases the pressure distribution showed strong periodicity, attributed to interaction between the grid systems. The initial spikes on the keel, for all cases with half-siding, were found to be close to the acoustic pressure limit,
. pV0 c. The pressure data could be smoothed by repeated local averaging, but this destroyed the peak structure of the distribution.
Total vertical force was smoothed by local averaging of the cumulative impulse a number of times, and then differentiating the result.
-IJJ a: ~ 0 C/) C/) UJ a: Q.
Vo = 10 FT/s
"'Y'\
I~ ~~ ")/ I~
..;;;:~
......... _,
0 GIRTH (FT)
.0
~ -:::... ~ .47
24
Figure 7 Impact Pressure Distributions for a 20 Degree Truncat.ed Wedged
9
4.5 Surge and Section Pynamics
The parameters for a simple wedge during impact are shown in Figure 8. Surge is conveniently represented by the ratio J.!=i/x. Since the water surface was free to deform in the simulations, surge could be
~t-x I
f--x----.~ I
.1/ WATER SURFACE WITH SURGE
-----------1 ..
Figure 8 Impact Geometry
readily estimated as illustrated by Figure 9 (the arrows in the figure indicate magnitude and velocity direction in the field). The value of J.1
observed was initially high but rapidly settled to a near constant value. These estimates of J.1 for the wedge cases are compared in Figure 10 with the theoretical approximation of Geers et al1 ~ and with two theoretical models of Chu and Abramson15
• Good correlation with the flat plate model can be seen, although selected experimental data cited in Reference 15 are
consistently lower than the theoretical models at small deadrise angles.
The surge ratio has some significance for impact force as shown by
Figure 11. The curves identified as 'R-K simulation' were obtained by integrating the equations of motion for a wedge using a Runge-Kutta method. Two values of J.1 were used in these calculations: the simulation estimate for this particular case, 1.69; and the average of estimates for all 20 degree wedge simulations, 1.77. The resulting 5 percent increment in surge ratio produces a 5 percent increase in vertical force. The third curve is the result of a PISCES simulation with 50 smoothing iterations applied. Even this amount of smoothing does not eliminate discretisation noise entirely.
10
I I \ \ \ \ \ \ \ \
I I I\\ I I \ \ \
I I I I I I I I I I
Figure 9 Frigate Section 3 after 0.367 Seconds: Initial Velocity 10 FPS
e SIMPLE WEDGES
0 TRUNCATED WEDGES
1.8
~ I
~_!___-I· _.!.~~2EE:.. -- ·=s -- 0 Q -- ~ ----- CHU AND ABRAMSON --ELLiPS£---
MODEL -
0 ..... < a: 1&.1 1.4 (.!) a: :::::1 Cf)
j 1.2
0 10 40 50
Figure 10 Surge Ratio for Simple and Truncated Wedges
11
40r-----,------,------~----~----~
-g Q 30 ' (/) ~ -
0 .I
PISCES, 50 SMOOTHING ITERATIONS
.2 .3 .4 .5 TIME ( s)
Figure 11 Time Histor~es of Vertical Force for a 20 Degree Simple Wedge with Initial Velocity 10 FPS
4.6 Impact Pressure Distribution
Although, as previously noted, the discretisation employed in these simulations was too coarse to adequately distinguish sharp peaks, the pressure signatures and the maxima occurring after the initial impact were in general fairly well represented, even after a number of smoothing iterations. To better determine the initial peaks, finer grids and shorter time steps could be used, but at much greater computational cost.
A typical set of pressure distributions for a t.runcated wedge was presented in Figure 7. Very little difference between simple and truncated wedges was observed. Some difference was seen in the initial peaks but these were heavily contaminated with noise; after a few time steps the distributions were virtually indistinguishable. After initial impact, most wedge cases approximated the classic Wagner16 pressure distribution, being fairly uniform in the region of the keel, rising to a peak (the amplitude of which depended on smoothing) at the waterline.
Pressure distributions for the frigate sections showed different characteristics as typified by Figure 12. Although initially like the wedge pressure distributions, the keel pressure falls more slowly with time while the waterline peak rapidly disappears. This results in the developed
12
:.•
~
•\ ..
pressure distribution having a triangular appearance as clearly shown in the figure. The wedges with highest deadrise, 40 degrees, exhibited intermediate behaviour, with the change from a Wagner pressure distribution
--~-- --t o~t-ri-angu-tar -be:i:ng--c-ons:i:derabl:y-d-e-layed~c-ompared-- w:i:th-the -fri-ga-te-sections-.----
-~20 -LLJ a: ::I (/) .0 (/) LLJ a: a.
Figure 12 Impact Pressure Distributions for Frigate Section 4
4.7 Fora Factor
Because of discretisation noise, the pressure distribution alone
would not give a good estimate of form factor, k, in the basic impact equation:
PMAX = (p/2)kv2 (11)
A number of alternative methods to evaluate form factor were tried; none were entirely satisfactory for frigate sections. One of the better methods for wedges consisted of obtaining a linear least-squares unbiased estimate of k from the total vertical force produced by a modified Wagner pressure distribution. If k is constant, not only should the estimate be asymptotic in time, but also uncertainty is progressively reduced. This was found to be so for most wedge cases, as illustrated by
13
Figure 13. As shown in the figure, after some initial excursions, the estimate stayed at a plateau for a while before falling slowly in value. This fall coincides with the decay of the Wagner pressure distribution.
Form factor, taken to be the aforementioned plateau value, is plotted versus wedge deadrise in Figure 14. The curves plotted for comparison are: Wagner- the classical result, Drop Tests- equation (5), and SC-OM - Figure 3.
40~--------~-----------r----------~--------~
30
k 20
10
0
!\ , . . ' I • . ' ' . . '
25
SMOOTHING ITERATIONS
-·-· 0 25
----- 100
50 TIME STEPS
75 100
Figure 13 Convergence of Form Factor Estimate for Simple 20 Degree Simple Wedge with Initial Velocity 10 FPS
14
,;.I
-~
80~-------r------~~-------r--------~------~
70
60
50
k 40
30
20
10
0
\ I
\ '
DROP\ TESTS
' \ ' \
\
', \ \. \ . ' '
' .......
e SIMPLE WEDGES
G> TRUNCATED WEDGES
sc-oM , }
,,~
' .. '~--...... ~
........... ------------~-:.-
10 20 30 40 50
Figure 14 Form Factor for Simple and Truncated Wedges
15
5. EXPERIMENTAL WORK
The experimental program was preceeded by an extensive review of the literature. In all published sources, there is serious discrepancy between the results of the various types of slamming investigations. Reconciling three-dimensional drop and seakeeping test results is at best difficult, so an experimental program was derived that would show comparison of the results of various techniques for the three hull forms given in Section 3.
As planned, the test program's main objectives were:
- to determine the form factor (or k-values) at each instrumented forward station;
- to compare the measured form factors with those predicted by the Ochi-Motter (OM) and the Stavovy-Chuang (SC) methods;
- to examine the effects of hull geometry on slam pressures;
- to develop a method for calculating the spatial and temporal variation of bow slam force.
The significant parameters measured during the experiments were the peak pressure value, rise time (from zero to maximum), duration of pressure pulse, the time between pulses, and the relative velocity between the hull and the water surface at the point of contact.
Three basic drop test configurations were selected, based on Schenzles' work8 :
- even keel free drops 1n calm water,
- pivoted (rotational) drops in calm water, and
- pivoted (rotational) drops in regular waves;
as shown in Figure 15. Additionally, seakeeping experiments in regular head waves were included in the program.
5.1 Drop Test Programme
The drop tests were carried out in a 60 ft long basin with a plunger-type wave maker at one end and a sloping gravel beach on the other. A modified towing carriage capable of accommodating the drop test rig and achieving high forward speed was used for the tests. The basin had a 10 ft wide wave channel in its center with transparent walls at the test section.
16
(a) FREE (EVEN-KEEL) DROPS IN STILL WATER
u-~illlil! .. DROP HEIGHT h =--t
~ STILL WATER
...., t
.I ::::z
Figure l5 Drop Test Types
MWL
The three models shown in Figure 5 were fitted with twenty-six Kulite XTMS 1-190 series diaphragm-type pressure transducers effective in the range 0 to 50 psi. The gauges were installed along the keel on the centreline and on one side of the lower hull from the F.P. to station 7. Wherever possible three transducers were installed at positions less than 1/10 draft (0.1T) from the keel, where maximum pressures were expected. Above the 0.1T line, transducer locations were chosen to provide a reasonably accurate girthwise pressure profile, and a few were located near the load water line forward to measure flare slamming effects. The model was also fitted with a motions package consisting of accelerometers and attitude indicators.
The drop test program was designed to achieve a wide range of relative velocities within an acceptable time and cost. A total of 849 drop tests were performed with the three models. The vertical drop velocity was readily controlled by the drop height, which was restricted to a maximum of 12 inches to avoid complete submergence of the bow. For the same reason, the maximum initial trim angle in the pivoted tests was 14 degrees. Calm water drops were repeated three times. Drops in regular waves had no repeat runs due to the difficulty in repeatedly dropping the model at exactly the same position on the wave. Table I gives the ranges of the drop test variables.
17
TABLE I Ran~es of DroE Test Variables
Minimum Increment Maximum
Drop Heights 2" 2" 12 11
Initial Trim Angles 20 20 14° Approximate Forward Speeds 0 0.5m/s 3m/s
5.2 Seakeeping Test Programme
To complement the drop tests, ARCTEC conducted a limited number of seakeeping tests using the National Research Council (NRC) of Canada's facilities in Ottawa. Although the original plan included seakeeping experiments with all three models, experiments were carried out only with the V-form model due to data handling difficulties.
The model was self-propelled in regular head waves in NRC's ship model experiment tank, with the model-carriage connections allowing freedom in pitch, heave and surge only. The model was fitted with the same pressure transducer array as in the drop tests, together with accelerometers and·sonic probes to derive motions. Relative motions were measured by means of capacitance probes on the model surface. The capacitance relative motion probes proved to be unsuitable for the derivation of instantaneous relative velocities. Without relative velocity at impact, form factors could not be derived accurately.
The seakeeping experiments were performed in three groups. In the first case, Froude number and wave frequency were varied to establish slamming zones, Table II. Then, wave slope was varied in the severe slamming region as shown in Table III. Finally, the bracketed experiment in Table II was repeated eleven times to obtain a measure of the scatter in the results.
The 89 test runs were expected to generate about 46 000 pressure peaks, calling for an unreasonable analysis effort. As a result, the analysis was restricted to seven transducers, using a PDP-11 computer digitising at 20 000 samples per second in the region of slams. Even with analysis of only 7 of 24 possible transducers, 50 hours of computer time were required for digitising.
18
,,
TABLE II Slarmning Zone Parameter Space in Seakeeping Tests
wiL/g 1.5 1.65 1.80 1. 95 2.10 2.25 2.40 2.55 2.70 ~-~-------- -~------~~
Fn
L. 0. 20
0.25 X
., '1- 0.34 X X X
~severe 0.35 X X X X X slamming
zone 0.40 X X X X X
o. 45 X X X X X X X X X
0.50 X X X X X X X
0.55 X X X X (X) X X X X
lOOI;A/A 1.18 1.23 0.99 1.04 1.22 1.33 1.56 1.60 1.55
Table III Wave Slope Variation in Severe Slamming Zone
ur/LTi, 1.8 2.1 2.4
lOOI;A/A RANGE Fn
o. 35 0.99-2.05 1.20-1.94 1.48-2.32
~ 0.45 0.99-2.00 1. 23-1.88 1.53-2.48
o. 55 o. 93-1.4 7 1.22-1.77 1.56-2.29
19
(a) SHIP-WAVE VELOCITIES AT IMPACT
SHIP MOTION
SHIP VELOCITIES Vn , Vv
DIRECTION OF WAVE PROPAGATION
WATER VELOCITIES NORMAL a TANGENTIAL {WAVE SURFACE Von ,Vo 1
WAVE HEIGHT : H HORIZONTAL, VERTICAL u , v
WAVE LENGTH = X.
UNIT NORMAL
(b) SHIP CO-ORDINATES AND ANGLES
BODY PLAN ~ = DEADRISE ANGLE
SHEER PLAN a = BUTTOCK ANGLE
Figure 16 Slam Geometry Definition
20
.. \. ,.,
5.3 Results
s. 3. 1 Drop Tests
- -For toedrop tests, tne c lear-wanea~wave~charttiel allowed-u-s·e--of-a video camera to record impacts, so that impact velocity and form factor could be reliably determined. The keel slam pressure is defined as follows,
p = (p/2)k'V2
k'= k/(1 + tan B cos2a)
V = Vh (V0 t cos 0 + Von sin 0) sin (a + T) +
Vv (V0 n cos 0 - V0 t sin 0) cos (a - T)
where the angles are defined ~n Figure 16 and the nomenclature.
.5 .a
--- 95% CONFIDENCE LINE
U- FORM STN 3, KEEL I ~~ I("' NORMAL FORM
STN 5, KEEL
I o· I :: •
I r· ~
I ~
I o~ I ::;
I ·t. ;/ .. =1i
• I • y· .. .;. I ·~
I • -t: I ,:;,.
I 1',~ • • I .; •• ·•
• I • a.; I
• !.j.-y .
I • • 1: ,. •
2 3 I 2 3 2 VELOCITY { FT /s)
3 5 8 10
(12)
(13)
(14)
Figure 17 Typical Pressure Plots from Stillwater Pivoted Drops
21
20
k
Typical results of the drop tests are shown in Figure 17 as pressure-velocity plots, together with derived dimensional form factors. It was apparent that 2 is not necessarily the most suitable index of velocity for predicting pressure, but this value was used in the conventional manner. In reality the index varied from 1.5 to 2.5, but this variation may be safely ignored considering that 95 percent confidence· limits for the derived coefficients encompass most of the data. Indeed, for design use, the upper 95th percentile limits, shown in Figure 17, could be considered more suitable predictors than the mean lines, although the mean value is appropriate for use of equation (10) for the prediction of design pressures.
50~------~------~~-----r------~------~------~r-------~--~--~----~
0 10
0 ..
SC-OM
--------- a FLAT CALM WATER DROPS
----{:
SC/COs4{3
- ... Q
PIVOTED CALM WATER DROPS
PIVOTED DROPS IN WAVES
G
G
............... ------· ---' ~ -~~-~-~
N --------__;_....._ ----. -----~--.
zo 30 eo 90
Figure 18 Experimental Drop Test Form Factors for Keel
22
Figure 18 shows a comparison of form factors derived from the drop tests, together with the SC-QM line first shown in Figure 3. As anticipated, there are differences between the flat and pivoted drop form factors, the former giving clearly larger pressures, except at low deadrise angl-e~Tlre-h~e-avy sc-att~e-r-an·d-reduced -pres-su re·s-seen-a-t-deadr-i-se-ang-le-s--
below 30 degrees could be due to air entrapment, but it is hard to credit
this at deadrise angles above five degrees. The pivoted drop results run into the upper, Ochi~otter, portion of the SC-QM line in Figure 18, but lie nearly an order of magnitude higher than the lower, Stavovy-Chuang portion. A plausable explanation for this disconcerting behaviour is
available.
Schmitke9 produced the composite SC-QM line by joining the two predictions where they crossed. Unfortunately, whereas Ochi-Motter predict
k, Stavovy-Chuang predict k111 •17
, where
The chain-dotted line in Figure 18 shows that correction of the lower portion of the SC-QM by the cos~ 8 divisor would improve agreement considerably.
Figure 18 also shows that keel form factor results for normal, U
and V-form models collapse to a single line when plotted to a base of 8, implying that hull form, per se, is not so important as deadrise angle.
This is the conventional assumption.
30 30 (/)
:E IJJ >- ..J
~20 s 20 c:S IJJ
b ::> ~
0 0 0
LLJ ' 0
b 0
10 10 0 ~ 0 0
0 a.. 0 0
0 10 20 30 0 10 20 30
PIVOTED, CALM FLAT,CALM
Figure 19 Comparison of Form Factor from Different Drop Test Types
Figure 19 compares derived form factors at individual stations to indicate differences between test techniques. Even in the reduced pressure zone (see Figure 18, 8<25°) pivoted drops in waves and in calm water yield consistent form factors. Conversely, comparison of the pivoted calm and flat calm results does not even suggest a relationship between the results of the two techniques. Perhaps this should be an expected result given the level of scatter in the flat calm line (Figure 18) at all deadrise angles. The value of flat drops in calm water must be questioned.
23
Velocity of Pressure
.e
LoJ•7 (.)
~.6 a:; ;:, 8-5 0 !5.4 >-~-3 UJ
~··t "-.1
0
Lo.l·7 (.) z ~-6 ;:, 8-5 0
~:4 >-~.3 Lo.l 5.2 Lo.l a:: "- .I
.0
Pulse Along Keel Model No. 266
PIVOTED DROPS IN WAVES
27 SAMPLES
z MEAN = 77.56 FT/s Ql SIGMA = 35. 60 FT Is tl 51 ~I C.. I ~I Ol
I
~ e2:t::Z: ,tz;t;?.
40 eo 120 ISO
VELOCITY ( FT /s)
Model No. 264 (Normal)
PIVOTED DROPS IN WAVES
49 SAMPLES
MEAN • 70.90 FT /s SIGMA = 39.13 FT/s
1
' l 200
Model No. 267 (Extreme V)
PIVOTED DROPS IN WAVES .a~~--~~~-r--~~--~--r--r--~
.7
0 40
57 SAMPLES
MEAN • 72.33 FT/s SIGMA= 44.85 FT/s
80 120 160 VELOCITY (FT/a)
Figure 20 Slam Pressure Pulse Velocities
24
(Extreme U)
)j
.,
Velocity of Pressure Pulse Along Girth Model No. 266 (Extreme U)
PIVOTED DROPS IN WAVES STATION 7K . 8r----r--.....---.--.---.---.----'-,.---.,..-.....---,.--,
~
~ .6 a: :l ~.5 0 !5.4 >-~.3 I&J
5.2 I&J a: Lo.. .I
0
.8 •
0
.8
~.7 z ~.6 :l u g.5 Lo.. 0_4 >-u ~.3 :l
fi:l.2 a: Lo..
0
~I ~I I&J(
B:t :=t 61
I I
20
· -&>-SAMPtES·· ---MEAN = 24.02 FT/s SIGMA= 10.83 FT/s
40 60 80 100 VELOCITY ( FT Is l
Model No. 264 (Normal)
PIVOTED DROPS IN WAVES, STATION 7K
~, ~I ~' a.. I :=I I
0
20
44 SAMPLES
MEAN = 29.27 FT Is SIGMA = 17.94 FT Is
40 60 80 VELOCITY ( FT/s)
100
Model No. 267 (Extreme V)
PIVOTED DROPS IN WAV'iS STATION 7K
z o, B, 81 11:( c.., ~, 01
I I
36 SAMPLES
MEAN • 30.39 FT /s SIGMA = 10.29 FT Is
40 60 80 100
VELOCITY (FT/sl
Figure 20 Slam Pressure Pulse Velocities cont'd
25
U)
Figure 20 shows histograms of slam pressure pulse velocity along both the keel and the girth for pivoted drops in waves. The average pressure pulse velocity is consistently higher than predicted by Ochi and Motter.
5.3.2 Seakeeping Tests
Failure of the seakeeping model relative motion gauges prevented the derivation of instantaneous relative velocities at impact, so it was necessary to use pitch and heave amplitudes and their average phases relative to the waves to estimate impact velocities. This was the source of considerable scatter in pressure velocity plots, so form factors were not derived; however, the seakeeping data shouid not be discounted.
It is revealing to superimpose seakeeping and pivoted drop results, as 1n Figure 21, for consistency of the two methods is clear; however, there is a disturbing amount of scatter at low pressures. There is also a suggestion that a common envelope encloses the two results; if so, the seakeeping tests might have benefitted from more extreme waveheight, notwithstanding difficulties with relative velocity measurement. More important, a statistical estimate of the upper envelope is potentially as useful a measure of pressure as the mean line used conventionally.
Figure 22 shows the poor repeatability of peak slamming pressures, with near order of magnitude variations. The variations, probably due to the typical characteristics of NRC's pneumatic wavemaker, would not be of importance had relative velocity been measured, but the absence of such measurements prevents the derivation of reliable form factors.
~ 20 • Q. 30
•- PIVOTED DROP TESTS o- SEAKEEING TESTS •••
...J I.IJ I.IJ ~
t:t
20
I.IJ 10
~ U) I.IJ a: Q. 5 ::E < ...J U)
~ Q.
0
STATION I
0
• 5
0
10
0 0
STATION 2
0
0
0 8
• ~· • •
<M •·• ... Cfi• ••
0 • • •
10 0 •
• • 0
.~ 00~
0 ~0 0
• ...
5 10 15. IMPACT VELOCITY (FT /s)
10
5
STATION 3 . ... ' . .. .... • • • • • • • • •
o• • :. • •• ~ 4• o.
0 8 0 ~
' i • 00 o.• 0 0 0 00 ooe
0 0 0 0 0 • s 0 C8)
5 10
Figure 21 V-Form Model Pressure-Velocity Relationships: Comparison of Pivoted Drop and Seakeeping Tests
26
~·
'·~~
<I
.,
12
::10 (/) a.. -a LU 0.: ::::>
6 (/) (/) ~ 0.: a.. 4 ::::!: <(
-' 2 (/)
0
POSITION 2K
r- -1 1
I I I
I I I I I I I I I I I I I I I I I I I I I I I I I
I I I I I
I I
I I I I I I I 1 I 1 I I I I I I
I I
I I I I I I I I I I I I I
I I I I 1 I I I I 1
I I I I 1 I I 1
I I r- I I
~ I I 1 ,, I I -PEAK '\ I ~- /r"\ \
/~ -~' ~ ~ ' ."'\ '~ " !"\ ~ ·:--: -~ -MEAN
" !"\ ~ :-- ~ ~ !:'-"' r"\
r"\
2 45 4647 4849 56 51 52 53 54 EXPT. N~ 9 !0 13 !7 !6 15 18 15 17 !4 !! #SLAMS
Figure 22 Repeatability of Slam Pressures During Seakeeping Tests
6. SEMULAXION/HODEL TEST COMPARISONS
Direct comparison between simulation and model test results is precluded by the noise content of the simulations and by different impact dynamics in each case. Nonetheless, this section highlights qualitative similarities observed in the development of the pressure signature around the girth.
A typical set of model test pressure traces is sketched in Figure 23. These data are at station 5 for a pivoted drop in calm water with an initial full scale impact velocity of approximately 20 fps. Pressures were recorded on the keel and at three locations A,B and C towards the waterline as shown in Figure 5. Different levels of ambient noise on each trace reflect different calibrations for each transducer.
The features seen in this figure are typical. 'The large oscillations on the keel trace may be due to air entrapment. At locations close to the keel, A and B, the pressure rise is almost instantaneous; at location C it is generally one to several orders of magnitude slower, depending upon the initial impact velocity. It is not possible, given the spatial resolution of these pressure measurements, to determine whether this pressure rise character change is gradual or abrupt. Also note that the initial peak remains reasonably constant between the keel and B, but is significantly reduced at C.
The dynamics of impact are different between a pivoted threedimensional drop and a simulated two-dimensional drop. Most notably, the times at which off-keel peaks are recorded in the example shown in Figure 23 suggest that the model is still accelerating after initial impact, whereas the equivalent two-dimensional simulation demonstrates a small deceleration almost immediately.
27
Despite the above, and ignoring the keel oscillations, the pressure traces in Figure 23 are consistent with the observations in Section 4.6 that the girthwise pressure distribution is initially Wagner-like, that is, peaked at the waterline, (at least to B). Further
around the girth, the pressure pulse becomes much more spread out resulting
in a more uniform pressure distribution, tending to triangular as the pulse disappears altogether. In Figure 24, data from drop 438 are plotted with
data from the corresponding simulation. The development of the pressure distribution as described above can be seen in both cases.
The transition from a Wagner to a triangular pressure distribution occurs as local deadrise is increa.sing, and may be associated with a critical value of deadrise. The simulations suggest that this transition is fairly abrupt in time, as shown in Figure 24 between step 72 and step 84. The value of deadrise for this transition, estimated from the simulation data, is shown in Figure 25. There is no systematic variation of transition position with impact velocity.
The width of the transition region in Figure 25 is due to numerical scatter in the simulations. Since the critical deadrise angle
for transition differs for each station in Figure 25, transition is not a function of deadrise angle alone, unlike form factor. In addition,
extrapolation suggests that for fuller sections, the critical value of deadrise may be quite low.
C/) 0 a. -LIJ 0: 2.0 ::;) C/)
~ 0 0: a. 1'*48 l ?' ~ ~ 20 •• I1U4t .._ ul on••,.J -'*•lias .....,_
j, c
0 ~~ ~~! ·~, ~~· ~"~~~~: :: 'l ~~lf :-1 0 50 100 150 2.00
MILLISECONDS FROM KEEL IMPACT
Figure 23 Drop 438,Sketch of Pressure Signatures at Station 5, to full scale
28
; .
0
40
0 40
__ .,__ DROP 438, FULL SCALE PISCES SIMULATION
'-------STEP 12/TIME = 16.2 m s
24/32.1
Figure 24 Comparison of Pressure Distribution Development
29
7.
70
MODEL 264
60
-0 50
UJ !:Q 0:: 40 a TRIANGULAR <t LIJ ' PRESSURE a 30 \ \_ DISTRIBUTIO ..J
5 g 20 WAGNER TRANSITION PRESSURE REGION
10 DISTRIBUTIONS
0 I 2 3 4 5 6 7 8 9 10 FULL.;SCALE GIRTH (FT)
Figure 25 Transition from Wagner to Triangular Girthwise Pressure Distribution from Simulation Data
CONCLUDIBG REMARKS
The work described in this memorandum has been aimed towards the eventual development of accurate slamming prediction capabilities for warship hulls. The initial tasks, presented here, consist of a comprehensive set of model experiments, undertaken by ARCTEC Canada Ltd., to provide a database of slamming data; and a series of numerical simulations to provide further insight into the fundamental processes of slamming and to aid with interpretation of the data. The simulations were
restricted by computational complexity to two dimensions.
Results from the numerical simulations agree well with two-dimensional theories and drop test data. The surge ratio,~' plays a significant role in two dimensional drop dynamics, although it is not clear to what extent this translates into three dimensions. Numerical noise made it necessary to smooth the results and made peak pressure determination uncertain.
A notable aspect of the simulation results was the transition of the developing bottom pressure distributio~·from a classic, Wagner, form to a uniform, or triangular, shape. This transition appears to be fairly abrupt, and may be associated with quite low local deadrise angles for· full sections, although it'is clearly not a function of deadrise alone.
30
.,...,
~,
j
'
'
(..,..
The transition was also noted on the higher deadrise angle wedge simulations, but was much delayed. Although spatial resolution of the bottom pressure was low, the experimental data appears to confirm the ex_~-~~~nce of this phe~~menon for three dimensional hul~_s.
In the experimental program, the results from the extensive drop test series and the more limited seakeeping experiments demonstrated that pivoted drop testing is a useful model experimental technique from which the slamming characteristics of warship hull forms can be evaluated. A good consistency was found between pivoted drops, both in calm water and 1n waves, and seakeeping tests.
The experimental results _show a fair degree of scatter, particularly at low impact velocities. These uncertainties are primarily in the measurement of impact velocity, which needs further attention in future work. Nonetheless, predictions of pressure-velocity relationships and the derived form factors are satisfactory for design applications.
A notable departure from previous theories is that the slam pressure pulse velocities observed for pivoted drops in waves were consistently higher than predicted by Ochi and Motter. Also, the previously adopted Stavovy Chuang-Ochi Motter line appears to underpredict form factors at deadrise angles greater than 25 degrees, apparently due to misinterpretation of the original Stavovy-Chuang data.
Consideration of the progress achieved this far leads to the following recommendations for further investigation:
- An improved technique for impact velocity measurement is required, possibly using some form of emergence probe.
- Specific experiments and numerical simulations should be conducted to clarify further some of the fundamental features of slamming, such as the spatial distribution and time history of pressures, particularly in three dimensions.
- A larger experimental database is required to allow the establishment of reliable empirical prediction methods.
8. ACKNOWLEDGEMENTS
The foundations of this research program were laid by Mr. R.T. Schmitke, formerly Leader, Ship Dynamics Group, DREA, now Head, Military Engineering Section, Defence Research Establishment Suffield. Mr. J.R. Gandier of CDC CYBERNET Services, Ottawa, performed the PISCES computer simulations.
31
REFERENCES
1. Mackay, M. and Schmitke, R.T.: 'PHHS, A FORTRAN Programme for Ship Pitch, Heave and Seakeeping Prediction', DREA Technical
. --Memor~ndum 7S/B, Apal 1978~ -- · -
2.
3.
Salvensen, N., Tuck, E.O., and Faltinsen, 0.: 'Ship Motions and Sea Loads', Trans. SNAME, Vol 78, 1970.
Schmitke, R.T.: 'Ship Sway, Roll and Yaw Motions in Oblique Seas', Trans. SNAME, Vol. 86, 1978.
4. Andrew, R.N._ and Lloyd, A.R.J.M.: 'Full-Scale Comparative Measurements of the Behaviour of Two Frigates in Severe Head Seas', Trans. RINA, Vol 123, 1981.
S. Ochi, M.K. and Motter, L.E.: 'Prediction of Slamming Characteristics and Hull Responses for Ship Design', Trans. SNAME, Vol. 81, 1973.
6. Chuang, S.L.: 'Slamming Tests of Three-Dimensional Models in Calm Water and Waves', NSRDC Report 4095, September 1973.
7. Ochi, M.K. and Bonilla-Norat, J.: 'Pressure-Velocity Relationship in Impact of a Ship Model Dropped onto the Water Surface and in Slamming in Waves', Naval Ship Research and Development Center Report 3153, 1970.
8. Schenzle, P., Broelmann, J., and Blume, P.: 'Systematische Untersuchung von Slamming - Drucken mit Hilfe von Dreh-Fall-Versuchen in Wellen', Institut flir Schiffbau, Universitat of Hamburg, Bericht Nr. 270, 1971.
9. Schmitke, R.T.: 'Improved Slamming Predictions for the PHHS Computer Program', DREA Technical Memorandum 79/A, February 1979.
10. Ochi, M.K. and Motter, L.E.: 1 A Method to Estimate Slamming Characteristics for Ship Design', Marine Technology, April 1971.
11. Stavovy, A.B. and Chuang, S.L.: 'Analytical Determination of Slamming Pressures for High-Speed Vehicles in Waves", J. Ship Research, Vol. 20, December 1976.
12. Schmitke, R.T. and Murdey, D.C.: 'Seakeeping and Resistance Trade-Offs in Frigate Hull Form Design', Proc. Thirteenth ONR, Tokyo, 1980.
DRrc-·-; ·H: tDING PAGE BLANK
33
13. Hancock, S.L.: 'Finite Difference Equations for PISCES 2DELK, A Coupled Euler Lagrange Continuum Mechanics Computer Program', Physics International Company Technical Memo TCAM 76-2, April 1976.
14. Geers, T.L., Loden, W.A. and Yee, H.C.: 'Boundary-Element Analysis of Fluid-Solid Impact', Winter Annual Meeting of the American Society of Mechanical Engineers, 1977, AMD-Vol. 26.
15.
16.
Chu, W.H. and Abramson, R.N.: 'Hydrodynamic Theories of Ship Slamming- Review and Extension', Journal of Ship Research, Vol. 4, No. 4, March 1961.
Wagner, H.: 'Uber-Stoss- und Gleitvorgange an der Oberflache von Flussigkeiten', Zeitschrift fur angewandte Mathematik und Mechanik, Vol. 12 No. 4, Berlin, August 1932.
17. Graham, R.: Private communication.
34
UNLIMITED DISTRIBUTION
UNCLASSIFIED !; ........ ., c ............. ..
,_
DOCUMENT CONTROL DATA - R o u -~rs..cu,w·-ctasiftaotion- of litfe.-bodV· of----trect end-indexing. ennotation _ _routt be 8f!!eted whet> the -•II document is clanifiedl
1. ORIGINATING ACTIVITY :Z.. DOCUMENT SECURITY CLASSIFICATION
111\Jr.J A~~TF'TF'n 2b. GROUP
J. DOCUMENT TITLE
Some Warship Slamming Investi$ations _,
4. DESCRIPTIVE NOTES (Type of report and inclusive datesl Technical Memorandum
5. AUTHOR lSI ll.nt name, first neme, middle initiall
Nethercote, W.C.E., Mackay, M, Menon, B. 6. DOCUM£I'I'I'OA~ January 1986 1a. ToTAL N~
1 OF PI.GES ,7b. !\!0. OF REFS l]
Sa. PROJECT OR GRANT NO. 9a. ORIGINATOR'S DOCUMENT NUMSERISI
DREA TECHNICAL MEMORANDUM 86/206
&b. CONTRACT NO. 9b. OTHER DOCUMENT HO.ISI (Any otMr nurnber1 tNt may be 8llivfted this docurnentl
10. DISTRIBUTION STATEMENT
Unlimited
11. SUPPLEMENTARY NOTES t2. SPONSORING ACTIVITY
13. ABSTRACT
Excessive slamming is the most common caus.e of speed reduction for frigates and destroyers in heavy head seas. DREA' s early interest in frigate/destroyer slamming was limited to the development of computer programs for ship seakeeping performance prediction. Published bottom slamming algorithms were adopted and refined to provide a slamming prediction capability for DREA software. In follow-on work described herein, both two-dimensional numerical simulations and three-dimensional model tests were performed by contractors to obtain better physical insight and expand the empirical data base. The two-dimensional numerical simulation results required considerable smoothing to reduce numerical noise, but still yielded form factors in satisfactory agreement with two-dimensional theoretical and experimental results~ The simulations also
:_ "'·
gave unique insight into the girthwise development of slamming pressures. The three-dimensiortal model tests indicated that pivoted drop tests in waves may be substituted for conventional seakeeping tests, but the general applicability of the results was limited by difficulties with measurement of relative impact velocity. Notwithstanding the velocity measurement difficulties, the three-dimensional tests provided important information on longitudinal and girthwise pressure pulse velocities. Further experimental work will be required to develop a usable experimental data base.
35
UN~SSIFIEn S.aorlty Cl .. slllcallon
KEY WOROS
warship frigate destroyer bottom slamming seakeeping
INSTRUCTIONS
1. ORIGINATING ACTIVITY: Enter the name end address of the nrganization inuing me clocument.
23. DOCUMENT SECURITY CLASSIFICATION: Enter the overall !ll!curity classification of the document includift(l ~ial wornino terms whenwer applicable.
2b. GROUP: Enter security reclassification group numbar. The thr" groupS are defined in Appendix "M' of the ORB Scurity Re;ulations.
3. DOCUMENT TITLE: Enter the complete document title in ell c.>Oital lenen. Tilles in all casos mould be unclauified. II •
sullicienttv descriptive title caftftOt be Mlected without ctanifi· cat•on. show title cl-ilication with me uaual one-apital-letter abbreviation in perenthesel immedlataly fotlowino the title.
4. DESCRIPTIVE NOTES: Enter the category of docvmetu, e.o. tnchnical rt11)01't, technical note or technical letter. If IICJCiropriate, enter the type of document, e.g. interim, 1)1'09fi!SS.
summ•y. annual or final. Give the indusive datn when a soecifh: reporting period Is covered.
5. AUTHORISI: Entllf the namelsl of authorlll as lhown on or in the document. Enter last name, first nama, middle initial. If military, show rank. The name of the principal author is en absolute minimum rttQUi,.,...t.
6. DOCUMENT DATE: Enter me date Cmonth, y..,) of Eltlblfthn:tent approval for publication of thl doCument.
7a. TOTAL NUMBER OF PAGES: The total paa- count lhOuid follow normal pagination procedures. i.e., enter 1M number
of CJ8911S containing information.
7b. NUMBER OF REFERENCES: Enter the totel numbar of reler.,.,cn cited in thlt doCument.
Ill. PROJECT OR GRANT NUMBER: If appropriate, enter thl Applicable research and d..,.topmant project or grant numbar under which the document was written.
8b. CONTRACT NUMBER: II appropriate, enter the applicable nuonber under which the document was written.
!l.l. ORIG:NATOR'S OOCUMENT NUMBEFICSI: Enter the r,fli•::at <lacument number by whiCh the document will be o<lf!f'lflli•.r! and controlled by the originating activity. This number must be unique to thit document.
36
9b. OTHER DOCUMENT NUMBER lSI: II thl d-t hat~ aalgned any other document numberS C either by the orifinator or by thl spontOr), also enter this nurnberhl.
10. DISTRIBUTION STATEMENT: Enter any llmitatiom on funtlw ·diuemiNition of the document, other thin those iltiDOied by -'tv ctenification, using stlndllfd stat-u such as:
111 "Qualified requasten INIY obtein copieS of this d-.ment from their delenc:8 documentation center."
121 "Annou-ent end diSMminatton of this dllc:ufMnt is not auti\Orind without prior approval from ciriginatino activity."
11. SUPPLEMENTARY NOTES: Use for additional explenltory nota.
12. SPONSORING ACTIVITY: Enter me nan" of the ctepartmantll protect office or leboratorv fii)Oft-ing the ,_ell and dlwlopment. Include addreu,
t 3. ABSTRACT: Enter an abstract giving a brief and lldllal 1Um1ft8tY of the document, even though It may also IlPPA' elslwtMwe in the bOdy of the document ltMif. It it highly desirable that the abstract of clauilied documents ba unclalsifled. Eldl peregrapll of thl abltntct 111111 end with an indication of the _,,;,., dlslilicetian of the lnfonnatlon in the 111'811'111h lunlfta the dOc:un.tt lttall is llftdluiliedl ,..,..__, 11 ITSl, lSI. ICl, IRI. or lUI.
Thl length of me al:iltract .,ouid be limited to 20 lifiiiHpaced
ltlndlrd typewrinen lines; 7"' incltas long.
t•. KEY WORDS: Key word& are technically meanift(lful termt or .,Ott llftr- that characterize a document end could ba helpful in catllogino· 1M document. Key -rds .,ould ba Mleeted 10 that no. IIIICUflty deasiflcetlon is reQUired. ldentiliwl, IUCII as aquipnwnt model designation, trade nerN, militlfY projlct code name, geographic loc:ation, may ba used 11 key woniJ but will ba lol'-<1 by 1ft indication of tachnical oonte•t.
,,
0.
) .. , . : /
.., 0?